Browser license plate detect + OCR

This project includes a client-side web page that runs license plate detection and OCR directly in the browser using onnxruntime-web.

Files

• web/index.html
• web/app.js
• web/models/yolo-v9-t-384-license-plates-end2end.onnx
• web/models/cct_xs_v1_global.onnx

Run

Serve the workspace as static files (required so the browser can fetch the ONNX file):

cd /workspaces/lic
python3 -m http.server 8000

Open:

• http://localhost:8000/web/

Usage

Option A (file upload):

1. Upload an image.
2. Optionally drag on the preview to crop to the plate region.
3. Click Run Detect + OCR.

Option B (webcam capture):

1. Click Start camera and allow camera access.
2. Click Capture frame to move the current webcam frame into the OCR canvas.
3. Optionally drag to crop.
4. Click Run Detect + OCR.

Inference runs entirely in the browser via JavaScript ONNX Runtime (wasm execution provider). The detector finds plate bounding boxes, then each crop is passed to OCR.